1. Field of the Invention
The present invention relates to acoustical transducer measurement methodology and apparatus utilizing efficient techniques for assessing their linear operational parameters.
2. Description of Prior Art
The use of any transducer usually involves a measurement of its operating parameters, principally its moving mass, support compliance, motional resistance(mechanical), electrical input impedance, force coupling factor and any other such parameter as required to determine its operational characteristics. This measurement is often done on prototype samples to determine if they meet design intent, on production units as a quality control measure, or for any number of other reasons. The specific unit being investigated will be called the Device Under Test or DUT.
Numerous techniques are in use today for measuring the transducers operational parameters. The most commonly used approach is probably the method (or a derivative of it) proposed by Neville Thiele in his landmark article on loudspeaker enclosure designs "Loudspeakers in Vented Boxes" which can be found in the Audio Engineering Society's Loudspeaker Antholog series. This paper is also published in the Proceedings of the IRE Australia, vol. 22, pp. 487-508.
The method of Thiele, which has come to be known as the Thiele-Small method and the parameters which are derived from it, the Thiele-Small parameters, are very useful in the design and utilization of electro-acoustic transducers, most notably, loudspeakers. (Small also published the use of Thiele's technique in his paper on "Closed-Box Loudspeaker Systems" also found in Loudspeaker Anthology.) Similar techniques can also be extended to the measurement of transducers used for hearing aids or other applications.
Once a sufficient set of operational parameters is known a relatively accurate prediction of the transducers performance can be determined. The requirement for this accurate prediction is usually that the frequency be low enough so that the wavelengths of the sound are greater than the size of the transducer itself. This is the so called "lumped parameter" region where simplified elements like resistors, inductors, capacitors, transformers, etc. can be used to represent the functionality of the real parameters, moving mass, support stiffness, motional resistance, force coupling factor and cone area, for example. The fundamental problem in determining the performance then becomes that of determining the values of the analogous components of resistance, inductance, etc. By measuring the readily available input impedance of the DUT and perturbing its mechanical system with either an added mass, an added compliance or some other perturbation, the operational parameters of the DUT can be calculated. This is basically the Thiele-Small method.
In many embodiments of the prior art the impedance of the circuit shown in FIG. 1 is measured. The system is perturbed and the impedance is measured again. From the shift in the resonance frequency which results from the know perturbation the moving mass and the support stiffness can be determined. This determination can be done by finding certain key frequencies such as resonance and the "half power points" of the motional impedance curve. The calculation of the values of electrical resistance (R.sub.e), electrical inductance (L.sub.e), force factor (Bl) cone mass (M.sub.m), support compliance (C.sub.m), motional resistance (R.sub.m), and radiating area (S.sub.d) is done by techniques shown in the prior art. These operational parameters can also be given in the analogous (Thiele-Small) parameters of resonance frequency (f.sub.s), electrical Q (Q.sub.e), mechanical Q (Q.sub.m), cone area (S.sub.d) and electrical resistance (R.sub.e). These two differing sets of operational parameters are completely equivalent being related to each other by a simple set of equations. It makes no difference which set is used in the derivation or analysis discussed in this application.
Phillips and Geddes, in "Efficient Loudspeaker Linear and Nonlinear Parameter Estimation" (presented at the Audio Engineering Society conference in October 1991, preprint #3164), described the use of the same lumped parameter circuit of Thiele but improve upon the measurement by "fitting" a complex curve to the measured data thereby utilizing more degrees of freedom in the analysis and reducing the expected error.
Jang and Kim disclose in their January 1994 JAES paper, vol.42, no. 1, "Identification of Loudspeaker nonlinearities Using the NARMAX Modeling Technique" the use of the input voltage signal and the cone motion, sensed via a laser, to determine the linear and nonlinear parameters. This was accomplished by means of a time domain fitting strategy which calculated the linear parameters with a low level input signal and the nonlinear parameters with a high level input signal. The input signal was Gaussian noise.
Easley et al. Disclose in AUTOMATED SYSTEM AND METHOD FOR AUTOMOTIVE AUDIO TESTING #5,361,305, a method for the rapid testing of an audio system in a vehicle. The system determines the functionality of the various channels and transducers but does not determine the transducer parameters.
Van Hout et al. disclose in METHOD AND DEVICE FOR TESTING FOR AUDIO INDUCED SYMPATHETIC BUZZES #5,491,753 the use of a standard test signal for diagnosing a buzz problem in the passenger compartment of a vehicle. No transducer parameters are determined.
Jeong and Ih teach in their April 1996 JAES paper, vol.44, No. 4, "Harmonic Balance Method for Estimating the Nonlinear Parameters of Electrodynamic Direct-Radiator Loudspeakers" that both the voltage and the current signal can be measured at one time thus eliminating the need for a perturbed mechanical system. A laser is used to directly detect the motion of the loudspeaker diaphragm. In this prior art discrete frequencies are used and the levels of the harmonics of these frequencies are determined. This data is used to calculate the linear and nonlinear parameters of the loudspeaker. The technique disclosed in this article is expensive to implement due to the use of the laser for output motion of the radiating surface. This method suffers from extremely complicated mathematical analysis.
Scott, Kelly and Leembruggen in "New Method of Characterizing Driver Linearity" in JAES Vol. 44, No. 4, April 1996 teach the use of a DC current to force the loudspeaker diaphragm off center in order to determine the linear and nonlinear parameters. By sensing the cones static displacement the force factor is determined. Once this is known the values of each of the other parameters can be calculated from the input impedance curve of the loudspeaker. This method suffers from the same set of drawbacks as the previous two, namely laborious, time consuming setup and expensive equipment.
Most recently, April 1997, Clark disclosed in "Precision Measurement of Loudspeaker Parameters" JAES, Vol. 45, no. 4, a straightforward direct measurement of the nonlinear loudspeaker parameters. This technique utilizes the same rear enclosure pressure method for diaphragm displacement as that used by Phillips and Geddes. Like other recent researchers Clark uses a laser as a displacement sensing device.
All of these prior art methods suffer from one or more of the following problems:
two measurements must be taken, which cannot be done simultaneously since either the mechanical system must be perturbed. The inherent assumption is that the system does not change from one measurement to the next (which is hard to impossible to control). Another key assumption is that when a parameter is perturbed, such as the cone mass by adding mass to it, or the compliance by placing the DUT in another box, or by loading the DUT with a different length of tubing such as used for hearing aid transducers, that this perturbation affects only that parameter intended to be effected. Experience has shown that this is seldom the case. Different values of perturbation nearly always lead to differing values of the derived operational parameters, a clear sign that more than one parameter is being perturbed. PA1 perturbation techniques usually require access to the elements of the DUT itself. For example to add mass to the cone requires access to the cone. The switching of gravity technique requires access to the cone to feedback the cone position into a servo control system or some other method of displacement sensing. It is often desirable to measure a transducer without the need to detach any of its attached parts, such as front grille or front cover, etc. It is also desirable to measure the parameters in-situ in order to determine the loudspeaker performance under real world conditions The perturbed load (changing rear volume) technique does not require access to the internal parts, but suffers from a mixed change (radiation mass and compliance) as well as cannot be done in-situ. PA1 expensive and cumbersome equipment is often required. It would be desirable to have a method which could utilize readily available Personal Computer (PC) "sound cards" or other PC data acquisition boards to acquire the data. This means that only simple measurements of voltage, current and sound pressure are possible. PA1 most techniques require multiple measurements and these measurements can be very time consuming. Any method used as a production quality control gauge must be extremely fast and require virtually no human intervention.